Forum Moderators: martinibuster
Thanks,
naitsirhc26
I've always felt it was keyword based. Probably account wide, if you have a smaller account.
So if sites are in related topics with similar keywords they will probably affect each other. The affect might be less if the sites are on totaly different topics.
I would imagine that people with quite large adsense accounts might not be smart-priced account wide, but it would make sense if us smaller publishers were.
As insane as this may sound we have some websites that are pr8 with thousands of people fighting to buy space and some that arent and when we have a smart pricing incident as happened this past October all of our properties drop by the exact same amount to bring our entire account down to a specific CPM.
It is account wide, we have exacting scientific data to support the case that smart pricing is account wide for all websites under one account.
A former moderator of this forum was told by a Google employee that it was account-based, as I recall. Things could have changed since then, of course.
As insane as this may sound we have some websites that are pr8 with thousands of people fighting to buy space and some that arent and when we have a smart pricing incident as happened this past October all of our properties drop by the exact same amount to bring our entire account down to a specific CPM.
According to Google, Smart pricing is about the anticipated likelihood of conversion. (PageRank is irrelevant.) Given the fact that AdSense is direct-response advertising, it's entirely reasonable that a site could be in demand by advertisers but be less than desirable from an AdSense point of view. Also, Google may feel that account-wide smart pricing helps to discourage the use of AdSense in ways that aren't good for advertisers. Or maybe it's a simple matter of conserving resources at Google's end: In an ideal world, it might make more sense to smart-price by individual page (remember Google's 2004 example of "a camera review" vs. "a page of photo tips"?), but in the real world, Google may feel that, with conversion rates from strong and weak pages averaging out, smart-pricing by account may be an acceptable compromise between statistical perfection and cost.
I've never seen a day where all the channels were smart priced. Most of the time, there are one or two that are NOT smart priced.
Unfortunately, the channels with the highest traffic always seem to get smart priced the most. Then that drags down the overall eCPM of the entire account.
Also, Google may feel that account-wide smart pricing helps to discourage the use of AdSense in ways that aren't good for advertisers.
Google is not a company run by idiots, that's for sure, so there is probably a sound reasoning for smart pricing whole accounts, even if it totally escapes me. However, if it is a form of punishment, that would be pretty stupid, since by punishing someone, you actually try to tell them they did something wrong and hope they change their ways, but smart pricing whole accounts is a very very, very poor form of communication, as opposed to say, actually telling people what and where the problem is. Affecting the whole account just muddies the water about the actual source of the problem.
You know, a very small percentage of publishers actually stay informed of the latest theories by reading forums like these. We aren't actually informed of the expected conversion rates of a site or told in any way that the use of adsense on a specific site isn't good for advertisers. They even officially tell people to put as many ads as possible and that low CTR on a section shouldn't be reason enough to remove ads from them as they may be converting nicely. How is it that a poorly performing site should have any effect on another, unrelated, better performing site, especially when we are all totally kept in the dark?
I'm not talking about obvious bad behaviors such as MFA sites or disguising ads as navigation menus, just a typical use of adsense on a typical site, but with a bad topic. On the surface, it just doesn't make any sense at all, and I just don't buy your arguments of trying to discourage behaviors or conserve resources or what not. It's either way smarter than that, or a brain fart on their end.
The fact that a totally honest site with say, poorly converting traffic, something like a myspace layout site targetting kids with no credit cards, should also punish another unrelated, but well converting, site about the best places to buy stamps is just ridiculous. What it does is tell people, if they are in such a position as to be able to decode the message, that they should remove the myspace layout site from their account. But how is the average joe going to understand that, especially when Google's official position on adsense is to put it virtually everywhere on the web and that statistical performance (CTR) isn't a factor? How are we to know what to do?
I hope you are wrong about this, otherwise, Google's shiny rep for me would drop a notch.
Any given increase on one site or sites results in a corresponding decrease on other sites to bring our entire account to within 99.99997% of its average cpm. This clearly shows that the entire account is being manipulated to make a predetermined maximum cpm payout for the account.
Thus the most likely reason they do not want you to have multiple accounts.
When our largest smartpricing incident happened in October the entire account dropped by a certain percentage, big sites and little sites with varied numbers flucuating on a daily basis but when averaged out the total payout per day equalled a exact same % drop for the entire account. This again proves there is such a thing as account wide smartpricing. It is real and it does happen, if you had seen our graphs you would agree.
I tend to think that on very large accounts this kicks in to handle the entire account as it is to complex and ensures a semi fair payout per cpm which is the standard model still within larger properties and networks.
Any given increase on one site or sites results in a corresponding decrease on other sites to bring our entire account to within 99.99997% of its average cpm. This clearly shows that the entire account is being manipulated to make a predetermined maximum cpm payout for the account.
This looks like not the SmartPricing per account phenomenon, but the capped earnings theory:
[webmasterworld.com...]
and more:
[google.com...]
[edit: oh, now I see your comment here:
[webmasterworld.com...]
]
Why doesn't somebody from Google respond to this thread? Or is it another trade secret we're not supposed to be talking about?
However, if it is a form of punishment, that would be pretty stupid, since by punishing someone, you actually try to tell them they did something wrong and hope they change their ways
Maybe Google is less interested in punishment than in natural selection, or in minimizing server overhead. I don't pretend to know why "smart pricing" is account-wide--assuming that it is account-wide--but I'd imagine there's a rationale behind whatever method is being used. Is the rationale correct? That's hard to say without knowing what it is.
There is no sound logic to automatic account-wide smart pricing. Why on earth would any business jump to the conclusion that if one site in one industry wasn't converting, then all sites in other industries can't and aren't converting well. It's absolutely ridiculous.
Is that how you'd run your business?
The only possible explanation for why Google once did this, if indeed it ever did, was due to lack of resources, i.e., it was unable to put enough engineers into producing code to separate one site conversion analysis from another, or that it just took up too much system resources to test one then another site.
Supposedly Google's a little smarter. It bases its smart pricing decisions on conversion tracking. So which conversion tracking? Only the worst site of an account?
How much money do you think it'll lose by jumping to the conclusion if one site in an account sucks, then all the others are junk, too? For every single publisher.
p/g
ie: you blow one account, no free lunch.
No sense in diversification, do as you like. Go out happy- work in your underpants. Ultimately you will be destroyed.
They even officially tell people to put as many ads as possible and that low CTR on a section shouldn't be reason enough to remove ads from them as they may be converting nicely
Don't worry, only the naive confuse a company's marketing message which is driven by its own agenda with real advice, if my own mother told me to put 3 ad units per page I won't listen, all this does eat away at the company's credibility.
The only possible explanation for why Google once did this, if indeed it ever did, was due to lack of resources, i.e., it was unable to put enough engineers into producing code to separate one site conversion analysis from another, or that it just took up too much system resources to test one then another site.
Or maybe things just average out, and it's more cost-effective and profitable not to be obsessive about calculating smart-pricing discounts for specific pages.
Also, if they aren't smart-pricing across the publisher's account, it would make more sense to smart-price for a given page or type of content, not for a given domain. Remember the examples of "a page of photo tips" (low conversions) and "a camera review" (higher conversions) that Google used when it introduced smart pricing? A site about photography might have both--along with forums, photo galleries, etc.--and the conversion rates between types of content on that site might vary more than the conversion rates between the domain and any other domains that the owner might have.
If smartpricing is indeed account wide then that would indeed be a problem in light of the first paragraph. Assume the first site is humming along and getting good clicks / conversions / traffic. The new sites that are being added would then drag down the first site?
If smartpricing is account wide this would say you have two options. 1 being to have multiple accounts (which I am led to believe is a no no) and 2. not place ads on sites until they are mature enough to not hurt the existing ones that convert well.
Definitely food for thought!
